SPADA Meeting Book

number of iterations required to discover a high-performing multiplex reaction. No modeling 537 algorithm is perfect, and there are many variables that are unknown even in the most 538 sophisticated design paradigm. However, using such sophisticated multiplex design should result 539 in many fewer design and experiment iterations and vastly superior assay performance.

540 541

5.0 Metrology for In Silico Analysis

542

Metrology is the science of measurement and serves an important, but often under- 543 appreciated role in the development and validation of in silico PCR assay design methods. 544 Measurement assurance concepts that help increase confidence and decrease uncertainty (the 545 error associated with a result) for experimental data (24) can also be applied to in silico 546 approaches.

547 548

549

5.1 Sources of measurement uncertainty

One of the key steps in PCR assay design is predicting the outcome of applying an assay to 550 one or more DNA templates. While the information needed to define a PCR assay depends on 551 the complexity of the computational model, relevant information can include: 1) primer and 552 probe oligo sequences and concentrations; 2) template sequences and concentrations; 3) salt 553 concentrations; 4) thermocycling times and temperatures; 5) nucleotide concentrations; 5) 554 polymerase concentration and properties (i.e. nucleotide extension rate); and 6) buffer 555 composition. All these parameters can affect the final outcome and therefore contribute to the 556 uncertainty, that is the error associated with the model prediction. 557 However, it is often challenging to obtain accurate, quantitative measurements for many of 558 these parameters in practice. Some information, like polymerase properties and the buffer 559

31